Trying to understand Node metrics


#1

Hi,
I’ve been looking into GET /pools/nodes API in Couchbase node (cb EE 5.5.3) on Linux OS, and it returns a json doc with lots of useful data. In this json, I’m just trying to understand the below section;
“hdd”:{
“total”:570948567040,
“quotaTotal”:570948567040,
“used”:12577400258,
“usedByData”:257629124,
“free”:28759392338
}
I presume, it is reporting Hard Disk usage stats of the data or index volume depending on the service it runs. But when I try to compare these figures (assuming those figures are in Bytes), with the df -h OS level stats, it does not match. I also considered Gigibyte and Gigabyte factors as well.
Could someone please clarify this? Is there a detailed APT doc available for /pools/nodes which explains each attribute of the json output?
Thanks in advance.


#2

Hi @bathige

You see the hdd stats under “storageTotals” which is hdd usage stats across the cluster. Are you comparing this with df -h stats from single node in cluster or the total?

Thanks,
Prathibha


#3

Hi Prathibha,
Thanks for your response.
I understand it is cluster wide stats but how it is calculated for MDS cluster model? Does it consider service type that each node runs or does it just consider all the nodes of the cluster ? Also what is the measurement of that figure? Is it Bytes?


#4

Hi,

  • The HDD measurements are in bytes and include stats for all active “data” nodes.
  • The usedByData field displays amount of storage used by the buckets + view indexes.
  • The other HDD fields such as “total” take into account data disk path as well as the index disk path that may be present on the data node.